Download Phase Transitions of Dirac Electrons Observed in Bismuth

Phase Transitions of Dirac Electrons in Bismuth.
July 2nd, 2008
In particle accelerators or in interstellar space, electrons are accelerated to velocities close
to the speed of light. At such high speeds (dubbed “relativistic”), their behavior must be
compatible with the general principles prescribed in Einstein’s theory of relativity. Is it
possible to observe the properties of “relativistic” electrons in a table-top experiment? A
recent experiment at Princeton explores this question by looking anew at electrons in the
element bismuth.
In metals, electrons move with a low velocity that barely gets above a few percent
the speed of light. At such low energies, the electrons are accurately described by the
Schrödinger equation. The existence of strong mutual repulsion (“interaction”) between
the electrons in metals, together with their high density, leads to a pot-pourri of exotic
collective electronic behavior at cryogenic temperatures, e.g. superconductivity,
magnetism, and the quantum Hall effect. Over the decades, detailed understanding of
these states has led to significant conceptual breakthroughs which have influenced many
subfields of physics. They have also led to important applications, e.g. the MRI machine
and ever-faster and smaller solid-state transistors.
At very high energies, the Schrödinger equation is supplanted by the Dirac
equation, which describes neutrinos (particles that travel at the speed of light), as well as
energetic electrons and protons. It would appear that the Dirac equation is never needed
in research in solids. However, it has long been known that, in the complicated
environment of a crystalline solid, an electron moves in the force field of a dense
concentration of ions. As a result, its motion can be dramatically altered. For example,
the mass of the electron is changed significantly. In a class of materials -- notably
graphene and bismuth -- the fundamental dispersion (how the electron’s energy grows
with its momentum) is changed qualitatively. The energy becomes linear in the
momentum, mimicking accurately that of neutrinos and relativistic electrons, despite the
slow velocity.
Lu Li (left) and Joseph Checkelsky and their faculty advisers report new findings on Dirac electrons in
bismuth. The figure displays a curve of the magnetic property of bismuth versus magnetic field. The inset
is a sketch of a model used to interpret the results.
In these solids, we need to employ the Dirac equation, but with the speed of light (a key
parameter in the equation) decreased 300-fold. In this “slow” universe, concepts familiar
in high-energy physics – “chirality” or handedness, anti-matter, and flavor symmetry –
suddenly become relevant. In addition, the crystal environment also introduces
conditions that are inaccessible in high-energy physics or astrophysics. The application
of a modest magnetic field profoundly alters the Dirac electrons’ behavior (to achieve
equivalent effects in vacuum, one would need intense magnetic fields found only on the
surface of a neutron star). The vast difference in scale introduces new electronic
phenomena that can be studied in the laboratory. A second important difference arises
from interaction effects. Whereas the electrons in ordinary metals are extremely efficient
in screening (masking) each other’s Coulomb force, Dirac electrons are much less so. As
a result, the latter are very sensitive to the mutual repulsion force, and susceptible to
electronic phase transitions into states in which the repulsion energy can be lowered. (A
familiar example of a phase transition is the melting of ice to water. In metals, electronic
transitions occur to the superconducting state or the ferromagnetic state as the
temperature is lowered.) Theorists have predicted that Dirac electrons with multiple
valleys or flavors should undergo a phase transition in a strong magnetic field.
A recent experiment by a Princeton team led by N. Phuan Ong (physics) and
Robert J. Cava (chemistry) involving graduate students Lu Li and Joseph Checkelsky
(photo) and postdoc Y. S. Hor, has obtained evidence for such a phase transition. In
bismuth, the Dirac electrons come in 3 flavors, represented by the 3 equivalent “valleys”
(energy minima) that may be occupied. When exposed to a magnetic field, the electrons
exert a weak torque on the sample. To tease out the torque signal (which is an excellent
diagnostic of the state of the electrons), Li and Checkelsky balanced a crystal of highpurity bismuth at the tip of a tiny gold cantilever, and measured the minute flexing of the
cantilever as the magnetic field is changed. At low magnetic fields, the detected
magnetic signal shows quantum oscillations which have long been studied in bismuth
(curve in figure). However, in high fields, the researchers observe abrupt jumps (arrows),
which signal phase transitions of the Dirac electrons to a new state. This is the first
observation of an electronic phase transition in a Group V element (the pnictides) in the
periodic table. The findings are scheduled to be published in Science July 25th, 2008.
The researchers adopt the following theory to interpret their findings. At low
magnetic fields, the Dirac electrons occupy equally the 3 equivalent valleys (shown as
ellipses in inset). At high magnetic field, all the electrons choose to occupy only one of
the valleys. Paradoxically, by the rules of quantum mechanics, this ensures that the
electrons spend most of the time avoiding each other (which lowers their total energy).
Another possibility is that the electrons occupy all 3 valleys as a “coherent
superposition”, in the parlance of physicists. In a tightly choreographed dance, all the
electrons spend “equal time” in the 3 valleys in precisely the same manner. This
situation, a hallmark of quantum mechanics, has no direct analog in everyday experience.
Future experiments will attempt to determine the correct candidate state.
In the quest to develop ever smaller and faster transistors, physicists and
engineers are attempting to harness the quantum behavior of electrons. A new promising
direction is to exploit the properties of these Dirac electrons. Research in bismuth and
graphene may uncover further new results that will dramatically expand the toolbox of
quantum researchers.